e2e: node: expose the running flag

Each e2e test knows it wants to restart a running kubelet or a
non-running kubelet. The vast majority of times, we want to
restart a running kubelet (e.g. to change config or to check
some properties hold across kubelet crashes/restarts), but sometimes
we stop the kubelet, do some actions and only then restart.

To accomodate both use cases, we just expose the `running` boolean
flag to the e2e tests.

Having the `restartKubelet` explicitly restarting a running kubelet
helps us to trobuleshoot e2e failures on which the kubelet
was supposed to be running, while it was not; attempting a restart
in such cases only murkied the waters further, making the
troubleshooting and the eventual fix harder.

In the happy path, no expected change in behaviour.

Signed-off-by: Francesco Romani <fromani@redhat.com>

test/e2e_node/device_plugin_test.go

@@ -220,7 +220,7 @@ func testDevicePlugin(f *framework.Framework, pluginSockDir string) {
 
 			restartTime := time.Now()
 			ginkgo.By("Restarting Kubelet")
-			restartKubelet()
+			restartKubelet(true)
 
 			// We need to wait for node to be ready before re-registering stub device plugin.
 			// Otherwise, Kubelet DeviceManager may remove the re-registered sockets after it starts.

test/e2e_node/gpu_device_plugin_test.go

@@ -103,7 +103,7 @@ var _ = SIGDescribe("NVIDIA GPU Device Plugin [Feature:GPUDevicePlugin][NodeFeat
 				f.PodClient().DeleteSync(p.Name, metav1.DeleteOptions{}, 2*time.Minute)
 			}
 
-			restartKubelet()
+			restartKubelet(true)
 
 			ginkgo.By("Waiting for GPUs to become unavailable on the local node")
 			gomega.Eventually(func() bool {
@@ -142,7 +142,7 @@ var _ = SIGDescribe("NVIDIA GPU Device Plugin [Feature:GPUDevicePlugin][NodeFeat
 			framework.ExpectEqual(devIDRestart1, devID1)
 
 			ginkgo.By("Restarting Kubelet")
-			restartKubelet()
+			restartKubelet(true)
 			framework.WaitForAllNodesSchedulable(f.ClientSet, 30*time.Minute)
 
 			ginkgo.By("Checking that scheduled pods can continue to run even after we delete device plugin and restart Kubelet.")
@@ -172,7 +172,7 @@ var _ = SIGDescribe("NVIDIA GPU Device Plugin [Feature:GPUDevicePlugin][NodeFeat
 			}
 
 			ginkgo.By("Restarting Kubelet")
-			restartKubelet()
+			restartKubelet(true)
 
 			ginkgo.By("Confirming that after a kubelet and pod restart, GPU assignment is kept")
 			ensurePodContainerRestart(f, p1.Name, p1.Name)
@@ -181,7 +181,7 @@ var _ = SIGDescribe("NVIDIA GPU Device Plugin [Feature:GPUDevicePlugin][NodeFeat
 
 			ginkgo.By("Restarting Kubelet and creating another pod")
 
-			restartKubelet()
+			restartKubelet(true)
 			framework.WaitForAllNodesSchedulable(f.ClientSet, 30*time.Minute)
 
 			ensurePodContainerRestart(f, p1.Name, p1.Name)

test/e2e_node/hugepages_test.go

@@ -207,7 +207,7 @@ var _ = SIGDescribe("HugePages [Serial] [Feature:HugePages][NodeSpecialFeature:H
 		framework.ExpectEqual(value.String(), "9Mi", "huge pages with size 3Mi should be supported")
 
 		ginkgo.By("restarting the node and verifying that huge pages with size 3Mi are not supported")
-		restartKubelet()
+		restartKubelet(true)
 
 		ginkgo.By("verifying that the hugepages-3Mi resource no longer is present")
 		gomega.Eventually(func() bool {

test/e2e_node/memory_manager_test.go

@@ -348,7 +348,7 @@ var _ = SIGDescribe("Memory Manager [Serial] [Feature:MemoryManager]", func() {
 				return kubeletHealthCheck(kubeletHealthCheckURL)
 			}, time.Minute, time.Second).Should(gomega.BeFalse())
 
-			restartKubelet()
+			restartKubelet(false)
 
 			// wait until the kubelet health check will pass
 			gomega.Eventually(func() bool {

test/e2e_node/podresources_test.go

@@ -731,7 +731,7 @@ var _ = SIGDescribe("POD Resources [Serial] [Feature:PodResources][NodeFeature:P
 			expectPodResources(1, cli, []podDesc{desc})
 
 			ginkgo.By("Restarting Kubelet")
-			restartKubelet()
+			restartKubelet(true)
 			framework.WaitForAllNodesSchedulable(f.ClientSet, framework.TestContext.NodeSchedulableTimeout)
 			expectPodResources(1, cli, []podDesc{desc})
 			tpd.deletePodsForTest(f)

test/e2e_node/util.go

@@ -414,8 +414,8 @@ func findKubletServiceName(running bool) string {
 	return kubeletServiceName
 }
 
-func restartKubelet() {
+func restartKubelet(running bool) {
-	kubeletServiceName := findKubletServiceName(false)
+	kubeletServiceName := findKubletServiceName(running)
 	// reset the kubelet service start-limit-hit
 	stdout, err := exec.Command("sudo", "systemctl", "reset-failed", kubeletServiceName).CombinedOutput()
 	framework.ExpectNoError(err, "Failed to reset kubelet start-limit-hit with systemctl: %v, %s", err, string(stdout))